Pulse or digital communications – Synchronizers – Phase displacement – slip or jitter correction
Reexamination Certificate
1999-07-13
2003-09-23
Pham, Chi (Department: 2631)
Pulse or digital communications
Synchronizers
Phase displacement, slip or jitter correction
C375S363000
Reexamination Certificate
active
06625241
ABSTRACT:
FIELD OF THE INVENTION
This invention relates to techniques for the transmitting serial data over a transmission medium. It also relates to methods for transmitting multiple bit streams on a single transmission medium. It further relates to Fibre Channel communications and communications protocol.
BACKGROUND OF THE INVENTION
The Fibre Channel (FC) standard provides a general transport vehicle for Upper Level Protocols such as Intelligent Peripheral Interface (IPI) and Small Computer System Interface (SCSI) command sets, the High-Performance Parallel Interface (HIPPI) data framing, IP (Internet Protocol), IEEE 802.2, and others. Proprietary and other command sets may also use and share Fibre Channel, but such use is not defined as part of the Fibre Channel standard. Logically, the Fibre Channel is a bidirectional point-to-point serial data channel, structured for high-performance capability. Physically, the Fibre Channel can be an interconnection of multiple communication points, called N_Ports, interconnected by a switching network, called a fabric, an arbitrated loop, or a point-to-point link. The word fibre is a general term used to cover all physical media types supported by the Fibre Channel, such as optical fiber, twisted pair, and coaxial cable.
The Fibre Channel standard specifies only a serial bit stream for transmission. Transfers between nodes over Fibre Channel occur between buffers. Information stored in a buffer (generally constructed from RAM) at a first node is sent from a transmitting port associated with that node, across a physical medium (i.e., the Fibre Channel), to a receiving port at a second node, and stored in a buffer at the second node. The basic unit of transfer for the contents of a buffer between two ports is the frame. A frame consists of a start-of-frame (SOF) word, a multi-word header, multiple data words, a cyclic redundancy check (CRC) word, and an end-of-frame (EOF) word.
Fibre Channel is structured as a set of hierarchical functions, each of which is described as a level. The lowest level, FC-0 (physical), has two components: interface and media. The media component defines the fibre, connectors and optical and electrical parameters for a variety of data rates. Coax and twisted pair versions are defined for limited distance applications. The interface component consists of transmitters, and receivers and their interfaces. The next level, FC-1 (transmission code and protocol), defines the transmission protocol which includes the serial encoding, decoding and error control. Level FC-2 (signaling protocol), which sits atop level FC-1, defines the signaling protocol which includes the frame structure and byte sequences. The next level, FC-3 (common services), defines a set of services which are common across multiple ports of a node. The highest level in the Fibre Channel standards set, FC-4 (mapping), defines the mapping between the lower levels of the Fibre Channel and the IPI and SCSI command sets, the HIPPI data framing, IP and other Upper Level Protocols (ULPs).
A buffer can be thought of as an ordered set of bytes numbered from 0 to n. Neither the actual length of a buffer nor the technology used to store the bytes are defined by the Fibre Channel standard. Stored bytes are transmitted in the order of increasing displacement (i.e., from low address to high address), starting with the first.
Fibre Channel does not provide for error correction of transmitted information. Instead, it relies solely on error detection and retransmission of inaccurately received information. Consequently, information stored in the buffer at the first node is not overwritten until it is determined that the information was accurately stored in the second node. The observed bit error rate (BER) over optical media seems to be about 1 error in 10E16 to 10E24 bits, which is well within the maximum 1E12 requirement of the Fibre Channel specification. With a BER of 1 error in 10E16 bits, and a Fibre Channel standard transmission rate of 1.0625 gigabaud/second, one error is expected on a single fibre of a link about once during each 1089 days. In order to provide for both the ordered sending of information bytes over the Fibre Channel and for the detection of errors, transmitted information bytes are encoded. Fibre Channel transmits information using an adaptive 8B/10B code. Code rules require that each 8-bit byte of data be transformed into a 10-bit Data Transmission Character. Two types of Transmission Characters are defined: Data and Special. The Special Transmission Characters are used to specify the maximum run length of a transmission and to provide word alignment.
The 8B/10B encoding scheme in Fibre Channel utilizes “running disparity” to detect most errors in received transmission characters. Cyclic redundancy checks (CRC) are used to detect errors which are undetected by running disparity. Running disparity is a requirement that the transmission code have a balance of ones and zeros over short periods of time. This requirement of balance necessitates a special encoding and decoding procedure. Some data bytes encode to transmission characters that have more ones than zeros; others have more zeros than ones; and still others have an equal number of ones and zeros. If a string of bytes were to encode to transmission characters where each transmission character has more ones than zeros, the transmission stream would quickly become unbalanced, resulting in the detection of an error at the receiving node. The 8B/10B algorithm used by Fibre Channel solves this problem by providing two encodings for each character having an unbalanced number of ones and zeros. For example, if a byte encodes to 011011 0101b, the first 6 bits are unbalanced, having 4 ones and 2 zeros. The complement, or alternate, encoding for the same data byte is 100100 0101b, which has 2 ones and 4 zeros in the first six bits. In order to maintain balance during transmission, each off-balance transmission character is always immediately followed by a character of opposite disparity. At the receiving node, the same balanced code rules apply to the decoding of transmission characters. It is illegal to decode a pattern of transmission characters that is unbalanced. Sixty-two percent of transmitted errors can be detected using the running disparity encoding scheme. Fibre Channel relies on CRC to detect the remaining thirty-eight percent.
The 8B/10B encoding scheme, in addition to facilitating the implementation of running disparity error detection, has the added advantage of maintaining transmission balance, whether it be light on/off balance for the loading of optical fiber or DC balance for the loading of AC-coupled copper media. Evenly-balanced code transmission facilitates receiver design.
The 8B/10B encoding scheme has only 390 valid patterns for transmission characters out of a total of 1024 possible patterns (2
10
). The number 390 is derived as follows: 256 byte patterns times two variations equals 512. However, 134 encoded patterns are fully balanced, so no alternate pattern is needed: 512−134=378. These 378 transmission characters are called data characters, or D-characters, for short. There are also twelve special characters, called K-characters, which are used for control functions, bringing the total to 390.
A unit consisting of four characters transmitted as a unit is called a transmission word, a total of 4×10, or 40, bits. A transmission word is the smallest complete transmission unit in Fibre Channel. The first of the four transmission characters can be either an encoded byte or a special character. The remaining three transmission characters are encoded bytes. Information transferred across Fibre Channel is not always an even multiple of four bytes. Consequently, the framing protocol has a provision to add pad, or filler, bytes to frames before transmission between nodes. The pad bytes are stripped as part of the framing protocol at the receiving node.
Special characters are used for signaling functions. The 8B/10B encoding algorithm guarantees that no data byte can be
Burd Kevin M
Pham Chi
LandOfFree
Data communications bit stream combiner/decombiner does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Data communications bit stream combiner/decombiner, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Data communications bit stream combiner/decombiner will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3091762